AMD Phenom X6 1090T Black Edition & 1055T launch on 4/27

[Lokinhow]

oh
Bad scores

[JumpingJack]

oh
Bad scores

Nah.... 32 bit, lower end X6 ... don't put much weight into this.... looks more like a really screwed up bench to me, and the voltages looks screwy too high to me. The CPUIDs are obviously idling down per C&Q, but the voltage is oddly high ... can't be right.

[accord99]

only 4.58x seems way too low

Turbo, the single thread 32-bit Cinebench score is more comparable to a 3GHz Phenom II.

[JumpingJack]

Turbo, the single thread 32-bit Cinebench score is more comparable to a 3GHz Phenom II.

Good point.

[vitchilo]

Who's the retard who tested a X6 at 800 mhz? What the hell?

[mindfury]

Who's the retard who tested a X6 at 800 mhz? What the hell?

It's just CNQ....Calm down...

[ajaidev]

lol those 1055T scores in cinebench look bad.

Does anyone know why were the tests done with 800Mhz or were these done with 2.8Ghz "The cinebench and vantage scores say otherwise" and if they were done in 800Mhz why ??

In cinebench the scores are like :

ST:

3246 - x4 920
4238 - i5 750
2758 - x6 1055T

MT:

11444 - x4 920
14142 - i5 750
12642 - x6 1055T

Is it possible that the 1055T is running at 2.8Ghz but the lack of cache is effecting the efficiency of the cores??

[crazydiamond]

yes 4.58x caught me off guard too, until i remembered turbo .... sidenote, higher cpu score in vantage than x4 @ 4150mhz cpu/2770mhz nb & lower score in cinebench multithread edit: these scores arent really relevant since they were on 64bit win7,,, just realized those scores were w/ 32bit :/

[FlanK3r]

Guys, this is only 32 bit Cinebench, AMD is more btter at x64 (better performance comparsion x86 vs x64 up than Intels a bit)

this score is simillary as x4 9xx Deneb core with 3.9 GHz
multi 4.6x is not bad! Man, do u think, it must be 5.9x?
Now we know score at x68, at x64 Thuban at 4 GHz will have about 4600x 4.62 = 21 000+(maybe more, case scaling mutli CPU is too better)

ajaidev : i5 750 score, what u posted is not x86, but x64. Know it at 100%. X4 965 is better in x64 than i5 750(turbo on) about 100 points.

[Oliverda]

Is it possible that the 1055T is running at 2.8Ghz but the lack of cache is effecting the efficiency of the cores??

It's very unlikely. Just think about the L3 cache-less Propus. IMHO something is wrong with these benches.

[RaV[666]]

Found a score for a istanbul on socket F.


Thats 64bit mind you.
For the people that dont know, istanbul is earlier version of lisbon that was server only ,D0 stepping, lisbon/thuban is E0.
So theres definietely something wrong with that score

That opteron is a six core 2.6Ghz no turbo on slow DDR2 cpu.So if anything, scores for Thuban will be higher.

[Manicdan]

good comparison
hopefully 1090T is good for 16k

[jcmarfilph]

I agree those posted Thuban results are crippled.

I'd like to see Cinebench R10 and R11.5 x64 and WPrime

[Oliverda]

I've just made a quick bench on Win 7 64bit (with 32bit cinebench exe):





@3.8GHz:




Cinebench

Thuban: 12642
Deneb: 9023

Vantage CPU (perf.)

Thuban: 14533
Deneb: 9570


So where is the problem?

[jcmarfilph]

Mem on the result posted is running at 3GB+ Unganged Mode @ 666Mhz @ 9-9-9-24. HyperMemory introduced some latencies too. ^_^

[informal]

Like many have stated before the scores are perfectly fine... C10 is a 32bit score and is like Oliverda showed 40% higher than on the same clocked Deneb,while scaling is off due to higher single tread score produced by Turbo Core functionality. Vantage score shows perfect scaling(50%).Oh and the CnQ is on obviously hence the 800Mhz CPUz shot. So,again where is the problem ?

[madcho]

I think the ram is slow.

And 3Go of ram. ... 32bit os with 4go of ram ... !!!

[Manicdan]

wouldnt 3GBs of ram mean that duel channel was off?

edit: NM it does have 4GBs but only 3.3 are avail. good job confusing me

[freeloader]

I wonder if we can disable Turbo mode? Strictly for Folding@Home I'd love to have all my cores running at the same speed.

[informal]

No,it means the test rig used 32bit OS... Ram is naturally dual ch. Again,the scores show higher single thread perf. due to Turbo and 40-50% higher multi thread. results over Deneb. Quite predictable if you ask me and very good .

I wonder if we can disable Turbo mode? Strictly for Folding@Home I'd love to have all my cores running at the same speed.

Turbo is for poorly threaded apps ,while folding your cores will run at the speed you set them to(be it stock or OC manually in BIOS or AoD).

[gOJDO]

Turbo is for noobz.

[Oliverda]

Turbo is for noobz.

IMHO turbo is a nice feature for the non-OCer users.

[haylui]

Like many have stated before the scores are perfectly fine... C10 is a 32bit score and is like Oliverda showed 40% higher than on the same clocked Deneb,while scaling is off due to higher single tread score produced by Turbo Core functionality. Vantage score shows perfect scaling(50%).Oh and the CnQ is on obviously hence the 800Mhz CPUz shot. So,again where is the problem ?

according to the link,
Phenom II x6 1055T 32bit Cinebench R10 render time is 1 min 9 sec, 69sec
Core i7 920 32bit Cinebench R10 render time also 69sec

Core i7 920 source: http://www.legitreviews.com/article/824/9/

edit: I guess logically, Phenom II x6 would have at least 5.3x of rendering time of a single core

[Oliverda]

according to the link,
Phenom II x6 1055T 32bit Cinebench R10 render time is 1 min 9 sec, 69sec
Core i7 920 32bit Cinebench R10 render time also 69sec

Core i7 920 source: http://www.legitreviews.com/article/824/9/

Add ~40% to Phenom II CPUs.

[haylui]

Add ~40% to X4 920.

????
I don't understand